The bank received the following data
Posted: Thu Jan 23, 2025 8:48 am
Counting external traffic (how many people passed by the office);
Heat maps of the location of employees and visitors in the hall (two independent maps);
Information about clients based on gender and age recognition;
Determining the type of emotion of a visitor upon entering the office and the emotional state of an employee, “measuring smiles”;
Recognition of consultant activity in the room (amount of time spent together with clients in the area); recognition of the presence of the director in the room, etc.
The accuracy of all model scenarios was up to 95-100%.
To legalize analytics, the bank independently notified russia telegram database visitors about video surveillance in branches and, if necessary, collected employee consent to process data.
The processing itself was carried out on the bank’s servers using predictive ML models developed by Beeline.
As a result, we implemented an interface and reporting system according to the customer's requirements, ensured the simultaneous operation of two detection models, and created a notification about the need to open an additional service window.
The first effect was an increase in customer satisfaction, which should lead to an increase in the bank’s turnover.
In similar cases of analyzing the behavior of employees in the retail sector, already in the first week after organizational conclusions, the number of clients left without attention decreases by 8%.
By managing the queue length, the conversion rate per customer increases by 5% and NPS by 7%. And that's not counting the additional information that the HR department receives.
Remember that if the system you invested in produces data, have all departments use it, not just one department. That way, the system will pay for itself faster.
Heat maps of the location of employees and visitors in the hall (two independent maps);
Information about clients based on gender and age recognition;
Determining the type of emotion of a visitor upon entering the office and the emotional state of an employee, “measuring smiles”;
Recognition of consultant activity in the room (amount of time spent together with clients in the area); recognition of the presence of the director in the room, etc.
The accuracy of all model scenarios was up to 95-100%.
To legalize analytics, the bank independently notified russia telegram database visitors about video surveillance in branches and, if necessary, collected employee consent to process data.
The processing itself was carried out on the bank’s servers using predictive ML models developed by Beeline.
As a result, we implemented an interface and reporting system according to the customer's requirements, ensured the simultaneous operation of two detection models, and created a notification about the need to open an additional service window.
The first effect was an increase in customer satisfaction, which should lead to an increase in the bank’s turnover.
In similar cases of analyzing the behavior of employees in the retail sector, already in the first week after organizational conclusions, the number of clients left without attention decreases by 8%.
By managing the queue length, the conversion rate per customer increases by 5% and NPS by 7%. And that's not counting the additional information that the HR department receives.
Remember that if the system you invested in produces data, have all departments use it, not just one department. That way, the system will pay for itself faster.